[Python] Remove dependency to cuQuantum and CUDA

Required prerequisites

• Search the issue tracker to check if your feature has already been mentioned or rejected in other issues.

Describe the feature

CUDA-Q is a platform for quantum computation, although the availability of a fast GPU-based simulator is of course an important aspect.
However, GPU is unnecessary when using NVQC or QPUs, and is problematic because of the time-consuming installation and the difficulty of making CUDA-Q dependent from other libraries.

In the case of C++, if you build without defining environment variables, you can install CUDA-Q without cuQuantum and CUDA, but in the case of Python, they are included in the dependencies e.g. https://github.com/NVIDIA/cuda-quantum/blob/main/pyproject.toml#L21, so they are forced to be installed.

There are two ways to solve this: make it optional or create a separate library like cudaq-core.

Start with the Python dependency declaration in pyproject.toml, especially the cuQuantum and CUDA-related entry referenced in the issue. Compare the optional-dependency and separate-library approaches described there, then verify that the Python package can be installed for NVQC or QPU use without requiring cuQuantum or CUDA.